Appendix - OLINK Data Analysis
Mixed Models - Test of Treatment effects
NPX - Summary statistics by TreatGroup
MIXED MODELS FOR ESTIMATING TREATGROUP EFFECTS
Given the substantial number of outcomes (NPXs of 184 proteins over 2 Olink Panels), mixed models were employed to screen the protein targets, enabling the differentiation of three treatment groups. \[ \begin{align*} NPX &= intercept + TreatGroup + Timepoint + TreatGroup:Timepoint + random(1|SUBJID) (Eq. 1)\\ NPX &= intercept + TreatGroup + AGE + Timepoint + TreatGroup:Timepoint:AGE + random(1|SUBJID) (Eq.2)\\ NPX &= intercept + TreatGroup + AGE + SEX + Timepoint + TreatGroup:Timepoint:AGE:SEX random(1|SUBJID) (Eq.3) \end{align*} \]
Scenario 1
For this scenario,
\[NPX = intercept + TreatGroup + Timepoint + TreatGroup:Timepoint + random(1|SUBJID)\] in which TreatGroup:Timepoint = Interaction between Treatment Group and Timepoint
Linear mixed models, test of coefficients
Post-hoc test
Scenario 2
For this scenario,
\[NPX = intercept + TreatGroup + AGE + Timepoint + TreatGroup:Timepoint:AGE + random(1|SUBJID)\] in which TreatGroup:Timepoint = Interaction between Treatment Group and Timepoint
Linear mixed models, test of coefficients
Post-hoc Test
Scenario 3
For this scenario,
\[NPX = intercept + TreatGroup + AGE + SEX + Timepoint + TreatGroup:Timepoint:AGE:SEX + random(1|SUBJID)\] in which TreatGroup:Timepoint = Interaction between Treatment Group and Timepoint